U.S. patent application number 13/122872 was filed with the patent office on 2011-10-27 for methods of manufacturing a mask blank substrate, a mask blank, a photomask, and a semiconductor device.
This patent application is currently assigned to HOYA CORPORATION. Invention is credited to Masaru Tanabe.
Application Number | 20110262846 13/122872 |
Document ID | / |
Family ID | 42780831 |
Filed Date | 2011-10-27 |
United States Patent
Application |
20110262846 |
Kind Code |
A1 |
Tanabe; Masaru |
October 27, 2011 |
METHODS OF MANUFACTURING A MASK BLANK SUBSTRATE, A MASK BLANK, A
PHOTOMASK, AND A SEMICONDUCTOR DEVICE
Abstract
A before-chucking main surface shape is measured in an actual
measurement region of a main surface of a substrate which has been
precision-polished and, based on the before-chucking main surface
shape of the substrate and a shape of a mask stage (1), an
after-chucking main surface shape of the substrate when a photomask
(2) manufactured from the substrate is set in an exposure apparatus
is obtained through simulation. A selection is made of the
substrate in which the after-chucking main surface shape has a
flatness of a first threshold value or less in a virtual
calculation region thereof. For the selected substrate, a
calculation is made of a first approximate curve approximate to a
cross-sectional shape along a first direction in a correction
region of the after-chucking main surface shape. Correction is
performed by calculating an approximate curved surface from the
first approximate curve and subtracting the approximate curved
surface from the after-chucking main surface shape to calculate an
after-correction main surface shape. A selection is made of the
substrate in which the after-correction main surface shape has a
flatness of a second threshold value or less in the correction
region.
Inventors: |
Tanabe; Masaru; (Tokyo,
JP) |
Assignee: |
HOYA CORPORATION
Shinjuku-ku, Tokyo
JP
|
Family ID: |
42780831 |
Appl. No.: |
13/122872 |
Filed: |
March 17, 2010 |
PCT Filed: |
March 17, 2010 |
PCT NO: |
PCT/JP2010/054511 |
371 Date: |
May 4, 2011 |
Current U.S.
Class: |
430/5 ; 430/319;
716/52 |
Current CPC
Class: |
G03F 1/60 20130101; G03F
1/82 20130101 |
Class at
Publication: |
430/5 ; 430/319;
716/52 |
International
Class: |
G03F 7/20 20060101
G03F007/20; G06F 17/50 20060101 G06F017/50; G03F 1/00 20060101
G03F001/00 |
Foreign Application Data
Date |
Code |
Application Number |
Mar 25, 2009 |
JP |
2009-074997 |
Claims
1. A method of manufacturing a mask blank substrate for use in a
photomask to be chucked on a mask stage of an exposure apparatus,
comprising: a step of preparing a substrate having a
precision-polished main surface, a step of measuring a
before-chucking main surface shape in an actual measurement region
of the main surface, a step of obtaining, through simulation, an
after-chucking main surface shape of the substrate when the
photomask manufactured from the substrate is set in the exposure
apparatus, based on the before-chucking main surface shape of the
substrate and a shape of the mask stage, a step of selecting the
substrate in which the after-chucking main surface shape has a
flatness of a first threshold value or less in a virtual
calculation region thereof, a step of performing correction for the
selected substrate by calculating a first approximate curve
approximate to a cross-sectional shape along a first direction in a
correction region of the after-chucking main surface shape,
calculating an approximate curved surface from the first
approximate curve, and subtracting the approximate curved surface
from the after-chucking main surface shape, thereby calculating an
after-correction main surface shape, and a step of selecting the
substrate in which the after-correction main surface shape has a
flatness of a second threshold value or less in the correction
region.
2. The method according to claim 1, wherein the step of calculating
the after-correction main surface shape performs correction by
calculating a second approximate curve approximate to a
cross-sectional shape along a second direction perpendicular to the
first direction, calculating an approximate curved surface from the
second approximate curve, and subtracting the approximate curved
surface, calculated from the second approximate curve, further from
the after-chucking main surface shape having been subjected to the
correction of subtracting the approximate curved surface calculated
from the first approximate curve.
3. A method of manufacturing a mask blank substrate for use in a
photomask to be chucked on a mask stage of an exposure apparatus,
comprising a step of preparing a substrate having a
precision-polished main surface, a step of measuring a
before-chucking main surface shape in an actual measurement region
of the main surface, a step of obtaining, through simulation, an
after-chucking main surface shape of the substrate when the
photomask manufactured from the substrate is set in the exposure
apparatus, based on the before-chucking main surface shape of the
substrate and a shape of the mask stage, a step of selecting the
substrate in which the after-chucking main surface shape has a
flatness of a first threshold value or less in a virtual
calculation region thereof, a step of performing correction for the
selected substrate by calculating a first approximate curve
approximate to a cross-sectional shape along a first direction in a
correction region of the after-chucking main surface shape,
calculating a second approximate curve approximate to a
cross-sectional shape along a second direction perpendicular to the
first direction, calculating an approximate curved surface from the
first approximate curve and the second approximate curve, and
subtracting the approximate curved surface from the after-chucking
main surface shape, thereby calculating an after-correction main
surface shape, and a step of selecting the substrate in which the
after-correction main surface shape has a flatness of a second
threshold value or less in the correction region.
4. The method according to claim 1, wherein the first approximate
curve is a quadratic curve or a quartic curve.
5. The method according to claim 2, wherein the second approximate
curve is a quadratic curve or a quartic curve.
6. The method according to claim 1, wherein the first threshold
value is 0.32 .mu.m and the second threshold value is 0.16
.mu.m.
7. The method according to claim 1, wherein the first threshold
value is 0.24 .mu.m and the second threshold value is 0.08
.mu.m.
8. The method according to claim 1, wherein the actual measurement
region is a region including the virtual calculation region and the
correction region and the virtual calculation region is a region
including the correction region.
9. The method according to claim 1, wherein the virtual calculation
region is a 142 mm square region.
10. The method according to claim 1, wherein the correction region
is a 132 mm square region.
11. The method according to claim 1, comprising a step of selecting
the substrate in which the before-chucking main surface shape has a
flatness of 0.4 .mu.m or less in an actual calculation region
thereof.
12. The method according to claim 11, wherein the actual
calculation region is a region including the virtual calculation
region and the correction region.
13. The method according to claim 11, wherein the actual
calculation region is a 142 mm square region.
14. The method according to claim 1, wherein the exposure apparatus
irradiates exposure light to the photomask through a slit which is
movable in the first direction and extends in the second
direction.
15. A method of manufacturing a mask blank, comprising forming a
thin film on the main surface, on the side where the
before-chucking main surface shape was measured, of the mask blank
substrate obtained by the method according to claim 1.
16. A method of manufacturing a photomask, comprising forming a
transfer pattern in the thin film of the mask blank obtained by the
method according to claim 15.
17. A method of manufacturing a semiconductor device, comprising a
step of chucking the photomask obtained by the method according to
claim 16 on a mask stage of an exposure apparatus which is capable
of performing main surface shape correction, and exposing and
transferring the pattern of the photomask to a resist film of a
wafer.
18. The method according to claim 3, wherein the first approximate
curve is a quadratic curve or a quartic curve.
19. The method according to claim 3, wherein the second approximate
curve is a quadratic curve or a quartic curve.
20. The method according to claim 3, wherein the first threshold
value is 0.32 .mu.m and the second threshold value is 0.16
.mu.m.
21. The method according to claim 3, wherein the first threshold
value is 0.24 .mu.m and the second threshold value is 0.08
.mu.m.
22. The method according to claim 3, wherein the actual measurement
region is a region including the virtual calculation region and the
correction region and the virtual calculation region is a region
including the correction region.
23. The method according to claim 3, wherein the virtual
calculation region is a 142 mm square region.
24. The method according to claim 3, wherein the correction region
is a 132 mm square region.
25. The method according to claim 3, comprising a step of selecting
the substrate in which the before-chucking main surface shape has a
flatness of 0.4 .mu.m or less in an actual calculation region
thereof.
26. The method according to claim 25, wherein the actual
calculation region is a region including the virtual calculation
region and the correction region.
27. The method according to claim 25, wherein the actual
calculation region is a 142 mm square region.
28. The method according to claim 3, wherein the exposure apparatus
irradiates exposure light to the photomask through a slit which is
movable in the first direction and extends in the second
direction.
29. A method of manufacturing a mask blank, comprising forming a
thin film on the main surface, on the side where the
before-chucking main surface shape was measured, of the mask blank
substrate obtained by the method according to claim 3.
30. A method of manufacturing a photomask, comprising forming a
transfer pattern in the thin film of the mask blank obtained by the
method according to claim 29.
31. A method of manufacturing a semiconductor device, comprising a
step of chucking the photomask obtained by the method according to
claim 30 on a mask stage of an exposure apparatus which is capable
of performing main surface shape correction, and exposing and
transferring the pattern of the photomask to a resist film of a
wafer.
Description
TECHNICAL FIELD
[0001] This invention relates to a method of manufacturing a mask
blank substrate for use in a mask blank which is for manufacturing
a photomask adapted to be used in a photolithography process.
BACKGROUND ART
[0002] In a photolithography process of semiconductor manufacturing
processes, a photomask is used. With the miniaturization of
semiconductor devices, a demand for miniaturization in this
photolithography process has been increasing. Particularly, an
increase in NA of an exposure apparatus using ArF exposure light
(193 nm) has proceeded for adaptation to the miniaturization and a
further increase in NA is proceeding following the introduction of
the immersion exposure technique. For adaptation to the demand for
the high miniaturization and the increase in NA described above, it
is required to enhance the flatness of a photomask. That is, in
view of the fact that the allowable width of position offset of a
transfer pattern due to the flatness has been reduced with the
reduction in pattern line width and that the focal depth in the
photolithography process has been reduced with the increase in NA,
the flatness of main surfaces of a mask substrate, particularly the
main surface on the side where a pattern is to be formed
(hereinafter, the main surface on this side will be referred to
simply as a main surface or a substrate main surface), is becoming
unignorable.
[0003] FIG. 6 is diagrams showing the shapes of a substrate of a
photomask before (before suction) and after (after suction) the
photomask is chucked in an exposure apparatus, wherein FIG. 6(a) is
a diagram showing the shape of the substrate before suction while
FIG. 6(b) shows the shape of the substrate after suction. As seen
from FIG. 6(a), four corners of the substrate are a little higher
than chuck areas of a main surface and the height gradually
increases toward its central portion. That is, generally circular
contour lines are shown in the substrate before suction. In the
substrate after suction, generally rectangular contour lines are
shown as seen from FIG. 6(b). In this manner, when the photomask is
chucked on a mask stage of the exposure apparatus by a vacuum
chuck, it may happen that the photomask is largely deformed upon
chucking due to the affinity with the mask stage or the vacuum
chuck.
[0004] Conventionally, since the product management is conducted in
terms of the flatness of the photomask before chucking, it may
happen that even if the photomask is excellent with its main
surface shape having high flatness before chucking, when the
photomask is chucked on the mask stage of the exposure apparatus,
the photomask is deformed depending on the affinity with the mask
stage or the vacuum chuck so that the flatness thereof is largely
degraded. This tendency is remarkable particularly in the case of a
substrate that tends to be distorted due to low symmetry of the
shape of its main surface. Thus, it is becoming necessary to
consider the flatness of the photomask when it is chucked by the
vacuum chuck. There has conventionally been proposed a method of
selecting a mask substrate having good flatness after chucking on a
mask stage of an exposure apparatus (see, e.g. Patent Document
1).
PRIOR ART DOCUMENT
Patent Document
[0005] Patent Document 1: JP-A-2003-50458
DISCLOSURE OF THE INVENTION
Problem to be Solved by the Invention
[0006] Using the mask substrate selection method of Patent Document
1, it is possible to relatively easily select a mask substrate
whose flatness after chucking becomes a predetermined level or
more. However, with the miniaturization of a transfer pattern, the
condition of flatness required for the shape of a mask substrate
after chucking is becoming more and more strict. For example, in
the case of a mask substrate (substrate for a mask blank) with a
152 mm square size, the flatness in a 132 mm square region is
required to be as high as 0.16 .mu.m or further 0.08 .mu.m. There
has been a problem that if a substrate with such a flatness is
selected by the selection method of Patent Document 1, the ratio of
successful products is largely reduced so that the production yield
is lowered.
[0007] On the other hand, with respect to the problem of the
substrate deformation of the photomask when it is chucked on the
mask stage, intensive studies have been made also on the exposure
apparatus supply side. As a result thereof, there has been
developed an exposure apparatus having a function of performing
correction in a height direction (cross-sectional direction of a
substrate) according to the shape of a photomask upon exposure. In
the case of a photomask for use in such an exposure apparatus that
can perform height-direction correction to the shape of the
photomask when it is chucked on a mask stage, there is room for
relaxation of a conventional mask blank substrate selection
criterion. If the selection criterion can be relaxed, the
production yield of mask blank substrates is improved. There has
been desired a method of selecting a mask blank substrate adapted
to the exposure apparatus having such a correction function.
[0008] This invention has been made in view of these circumstances
and has an object to provide a method of manufacturing a mask blank
substrate which is adapted to an exposure apparatus having a
function of performing height-direction correction to the shape of
a photomask.
Means for Solving the Problem
[0009] According to one aspect of this invention, there is provided
a method of manufacturing a mask blank substrate for use in a
photomask to be chucked on a mask stage of an exposure apparatus,
comprising a step of preparing a substrate having a
precision-polished main surface, a step of measuring a
before-chucking main surface shape in an actual measurement region
of the main surface, a step of obtaining, through simulation, an
after-chucking main surface shape of the substrate when the
photomask manufactured from the substrate is set in the exposure
apparatus, based on the before-chucking main surface shape of the
substrate and a shape of the mask stage, a step of selecting the
substrate in which the after-chucking main surface shape has a
flatness of a first threshold value or less in a virtual
calculation region thereof, a step of performing correction for the
selected substrate by calculating a first approximate curve
approximate to a cross-sectional shape along a first direction in a
correction region of the after-chucking main surface shape,
calculating an approximate curved surface from the first
approximate curve, and subtracting the approximate curved surface
from the after-chucking main surface shape, thereby calculating an
after-correction main surface shape, and a step of selecting the
substrate in which the after-correction main surface shape has a
flatness of a second threshold value or less in the correction
region.
[0010] According to this method, an after-chucking main surface
shape of a substrate (photomask) when a photomask is chucked on the
mask stage of the exposure apparatus is obtained through simulation
and then the same correction as substrate main surface shape
correction in a height direction (cross-sectional direction of the
substrate) which is actually performed by the exposure apparatus is
further performed to calculate a flatness of the photomask. As a
consequence, the ratio of successful mask blank substrates with an
after-correction flatness satisfying a reference flatness is
improved in unsuccessful mask blank substrates which do not satisfy
the reference flatness when the height-direction correction by the
exposure apparatus is not considered. Therefore, it is possible to
achieve significant improvement in the production yield.
[0011] According to another aspect of this invention, there is
provided a method of manufacturing a mask blank substrate for use
in a photomask to be chucked on a mask stage of an exposure
apparatus, comprising a step of preparing a substrate having a
precision-polished main surface, a step of measuring a
before-chucking main surface shape in an actual measurement region
of the main surface, a step of obtaining, through simulation, an
after-chucking main surface shape of the substrate when the
photomask manufactured from the substrate is set in the exposure
apparatus, based on the before-chucking main surface shape of the
substrate and a shape of the mask stage, a step of selecting the
substrate in which the after-chucking main surface shape has a
flatness of a first threshold value or less in a virtual
calculation region thereof, a step of performing correction for the
selected substrate by calculating a first approximate curve
approximate to a cross-sectional shape along a first direction in a
correction region of the after-chucking main surface shape,
calculating a second approximate curve approximate to a
cross-sectional shape along a second direction perpendicular to the
first direction, calculating an approximate curved surface from the
first approximate curve and the second approximate curve, and
subtracting the approximate curved surface from the after-chucking
main surface shape, thereby calculating an after-correction main
surface shape, and a step of selecting the substrate in which the
after-correction main surface shape has a flatness of a second
threshold value or less in the correction region.
[0012] According to this method, in the case where height-direction
correction by the exposure apparatus functions not only in one
direction (first direction), but also in a direction (second
direction) perpendicular to that one direction, the ratio of
successful mask blank substrates with an after-correction flatness
satisfying a reference flatness is further improved in unsuccessful
mask blank substrates which do not satisfy the reference flatness
when the height-direction correction by the exposure apparatus is
not considered.
[0013] When a photomask is chucked on the mask stage of the
exposure apparatus, a substrate tends to be deformed into a
quadratic surface due to a suction force of a chuck. Taking this
modification into account, in a process of polishing a photomask
blank substrate, the processing is carried out aiming at obtaining
a substrate main surface with a convex shape which is relatively
high at its central portion and relatively low at its peripheral
portion. However, the substrate often has a main surface shape with
a large quadratic component due to variation in polishing
processing accuracy, the processing which is carried out in a
manner to avoid a concave shape, and so on. As a consequence, it
often happens that the quadratic component remains in an
after-chucking main surface shape of the substrate after the
photomask is chucked on the mask stage so that the flatness does
not satisfy a predetermined value. In this case, using a quadratic
curve as a first approximate curve and/or a second approximate
curve to correct the quadratic component of the after-chucking main
surface shape, the flatness can be set to the predetermined level
or more and thus the ratio of substrates to be successful products
is further improved.
[0014] In the case of a substrate having a before-chucking main
surface shape with a strong quartic component (tendency to a
quartic surface is strong), since deformation with a tendency to a
quadratic surface is applied when a photomask is chucked on the
mask stage, an after-chucking main surface tends to have a shape in
which the quartic component remains (with a tendency to a quartic
surface). In this case, using a quartic curve as a first
approximate curve and/or a second approximate curve to correct the
quartic component of the after-chucking main surface shape, the
flatness can be set to a predetermined level or more and thus the
ratio of substrates to be successful products is further
improved.
[0015] If optical correction by the exposure apparatus is allowed
such that the flatness correction amount of the substrate main
surface using the first approximate curve and/or the second
approximate curve exceeds 0.16 .mu.m, the astigmatism becomes large
when a transfer pattern of the photomask is actually transferred to
a resist film on a semiconductor wafer by the exposure apparatus.
As a consequence, there is a possibility that the image formation
of the transfer pattern is degraded so that a predetermined or more
pattern resolution is not satisfied. Thus, this is not
preferable.
[0016] Since the deformation of the substrate is caused by chucking
of the photomask by the exposure apparatus, it is necessary to also
consider a region outside a region where a transfer pattern is to
be formed. In general, a region where a transfer pattern is formed
in a thin film of a photomask is often set to a 132 mm.times.104 mm
region and, if the flatness of a 132 mm square region is good,
there is often no problem. However, if the flatness of a region
outside thereof is poor, there is a possibility that the substrate
deformation amount before and after chucking is large. If the
substrate deformation amount is large, the displacement amount of
the transfer pattern formed on a substrate main surface is large so
that the pattern position accuracy is lowered. Taking them into
account, a virtual calculation region for selecting a substrate to
be simulated is preferably set to a 142 mm square region.
[0017] In general, the region where the transfer pattern is formed
in the thin film of the photomask is often set to the 132
mm.times.104 mm region. In order to allow the formation of the
transfer pattern in any direction, it is preferable to set a
correction region in the 132 mm square region and to ensure the
flatness of the substrate main surface after chucking within a
predetermined value.
[0018] It is not that the flatness of the before-chucking main
surface shape of the substrate can take any value as long as the
after-chucking main surface shape satisfies the first threshold
value. In the case of a substrate in which an after-chucking main
surface shape is good while the flatness of a before-chucking main
surface shape is poor, the substrate deformation amount before and
after chucking is large and thus the displacement amount of a
transfer pattern formed on a substrate main surface is also large
so that the pattern position accuracy is lowered. As a substrate to
be simulated, it is preferable to select a substrate with a
flatness of 0.4 .mu.m or less in an actual calculation region. The
actual calculation region is preferably a region including a
virtual calculation region as a region for calculating a flatness
from an after-chucking main surface shape after simulation and a
correction region as a region for performing correction of
subtracting an approximate curved surface. Further, taking into
account the measurement accuracy of a flatness measuring apparatus,
the accuracy of a simulation, and so on, the actual calculation
region is more preferably a 142 mm square region.
[0019] In a mask blank substrate manufacturing method according to
still another aspect of this invention, it is more preferable to
use an exposure apparatus adapted to irradiate exposure light to a
photomask through a slit which is movable in the first direction
and extends in the second direction.
[0020] According to still another aspect of this invention, there
is provided a method of manufacturing a mask blank, comprising
forming a thin film on the main surface, on the side where the
before-chucking main surface shape was measured, of the mask blank
substrate obtained by the above-mentioned method.
[0021] According to yet another aspect of this invention, there is
provided a method of manufacturing a photomask, comprising forming
a transfer pattern in the thin film of the mask blank obtained by
the above-mentioned method.
Effect of the Invention
[0022] According to a mask blank substrate manufacturing method of
this invention, an after-chucking main surface shape of a substrate
(photomask) when a photomask is chucked on a mask stage of an
exposure apparatus is obtained through simulation and then the same
correction as substrate main surface shape correction in a height
direction (cross-sectional direction of the substrate) which is
actually performed by the exposure apparatus is further performed
to calculate a flatness of the photomask. As a consequence, the
ratio of successful mask blank substrates with an after-correction
flatness satisfying a reference flatness is improved in
unsuccessful mask blank substrates which do not satisfy the
reference flatness when the height-direction correction by the
exposure apparatus is not considered. Therefore, it is possible to
achieve significant improvement in the production yield.
BRIEF DESCRIPTION OF THE DRAWINGS
[0023] FIG. 1 is diagrams for explaining part of an exposure
apparatus having a photomask shape correction function, wherein (a)
is a plan view and (b) is a side view.
[0024] FIG. 2 is a flowchart for explaining a mask blank substrate
manufacturing method according to an embodiment of this
invention.
[0025] FIG. 3 is diagrams for explaining main surface shape
correction in a scan direction, wherein (a) is a diagram showing
positions where cross-sectional shapes of a substrate are obtained
and (b) is a diagram showing the cross-sectional shapes of the
substrate.
[0026] FIG. 4 is diagrams for explaining main surface shape
correction in a slit direction, wherein (a) is a diagram showing
positions where cross-sectional shapes of a substrate are obtained
and (b) is a diagram showing the cross-sectional shapes of the
substrate.
[0027] FIG. 5 is a diagram showing a schematic structure of a
sputtering apparatus for use in manufacturing a mask blank
according to the embodiment of this invention.
[0028] FIG. 6 is diagrams showing the shapes of a substrate of a
photomask before and after the substrate is chucked in an exposure
apparatus, wherein (a) is a diagram showing a before-chucking main
surface shape and (b) is a diagram showing an after-chucking main
surface shape.
[0029] FIG. 7 is diagrams for explaining main surface shape
correction in Example 1, wherein (a) is a diagram showing positions
where cross-sectional shapes of a substrate are obtained, (b) is a
diagram showing the cross-sectional shapes of the substrate, (c) is
a diagram showing an approximate curved surface, and (d) is a
diagram showing an after-correction main surface shape.
[0030] FIG. 8 is diagrams for explaining main surface shape
correction in Example 2, wherein (a) is a diagram showing positions
where cross-sectional shapes of a substrate are obtained, (b) is a
diagram showing the cross-sectional shapes of the substrate, (c) is
a diagram showing an approximate curved surface, and (d) is a
diagram showing an after-correction main surface shape.
[0031] FIG. 9 is diagrams for explaining main surface shape
correction in Example 3, wherein (a) is a diagram showing an
approximate curved surface and (b) is a diagram showing an
after-correction main surface shape.
MODE FOR CARRYING OUT THE INVENTION
[0032] Hereinbelow, an embodiment of this invention will be
described in detail with reference to the drawings.
[0033] A mask blank substrate manufacturing method of this
invention is for obtaining a mask blank substrate for a photomask
that can be used in an exposure apparatus having a photomask shape
correction function. Herein, the exposure apparatus having the
photomask shape correction function will be described.
[0034] FIG. 1 is diagrams for explaining part of the exposure
apparatus having the photomask shape correction function, wherein
(a) is a plan view and (b) is a side view. In this exposure
apparatus, a photomask 2 is placed on a mask stage 1 and it is
chucked on the mask stage 1 by a chuck 1a. An illumination optical
system 5 and a slit member 3 having a slit 3a are disposed above
the mask stage 1 and a light source 4 is disposed above the slit
member 3. Further, a reduction optical system 6 and a semiconductor
wafer W placed on a wafer stage 7 are located below the mask stage
1.
[0035] In this exposure apparatus, exposure to the semiconductor
wafer W is carried out while moving the mask stage 1 chucking
thereon the photomask 2 in a scan direction and moving the wafer
stage 7 in a direction opposite to the moving direction of the mask
stage 1. In this event, light from the light source 4 passes
through the slit 3a so as to be irradiated to the photomask 2 and
then is irradiated to the semiconductor wafer W through the
photomask 2. As a consequence, a transfer pattern is exposed to a
photoresist provided on the semiconductor wafer W. The moving
direction (scan direction) of the mask stage 1 and an extending
direction (longitudinal direction) of the slit 3a are substantially
perpendicular to each other.
[0036] In this exposure apparatus, it is possible to perform main
surface shape correction according to the shape of the photomask
which is measured and obtained in advance. In the scan direction,
the main surface shape correction is performed by changing the
relative distance between the mask stage and the wafer stage 7
placing thereon the semiconductor wafer W to change a scan path. On
the other hand, in the slit direction, the main surface shape
correction may be performed by changing the astigmatism to change
the shape of illumination light. In the case of this exposure
apparatus, the description has been given of the type that can
perform the main surface shape correction in the two directions,
i.e. the scan direction and the slit direction. However, depending
on an exposure apparatus, main surface shape correction is
performed only in the scan direction or only in the slit
direction.
[0037] The mask blank substrate manufacturing method of this
invention is a method of manufacturing a mask blank substrate for
use in a photomask to be chucked on a mask stage of an exposure
apparatus, comprising a step of preparing a substrate having a
precision-polished main surface, a step of measuring a
before-chucking main surface shape in an actual measurement region
of the main surface, a step of obtaining, through simulation, an
after-chucking main surface shape of the substrate when the
photomask manufactured from the substrate is set in the exposure
apparatus, based on the before-chucking main surface shape of the
substrate and a shape of the mask stage, a step of selecting the
substrate in which the after-chucking main surface shape has a
flatness of a first threshold value or less in a virtual
calculation region thereof, a step of performing correction for the
selected substrate by calculating a first approximate curve
approximate to a cross-sectional shape along a first direction in a
correction region of the after-chucking main surface shape,
calculating an approximate curved surface from the first
approximate curve, and subtracting the approximate curved surface
from the after-chucking main surface shape, thereby calculating an
after-correction main surface shape, and a step of selecting the
substrate in which the after-correction main surface shape has a
flatness of a second threshold value or less in the correction
region.
[0038] Alternatively, the mask blank substrate manufacturing method
of this invention is a method of manufacturing a mask blank
substrate for use in a photomask to be chucked on a mask stage of
an exposure apparatus, comprising a step of preparing a substrate
having a precision-polished main surface, a step of measuring a
before-chucking main surface shape in an actual measurement region
of the main surface, a step of obtaining, through simulation, an
after-chucking main surface shape of the substrate when the
photomask manufactured from the substrate is set in the exposure
apparatus, based on the before-chucking main surface shape of the
substrate and a shape of the mask stage, a step of selecting the
substrate in which the after-chucking main surface shape has a
flatness of a first threshold value or less in a virtual
calculation region thereof, a step of performing correction for the
selected substrate by calculating a first approximate curve
approximate to a cross-sectional shape along a first direction in a
correction region of the after-chucking main surface shape,
calculating a second approximate curve approximate to a
cross-sectional shape along a second direction perpendicular to the
first direction, calculating an approximate curved surface from the
first approximate curve and the second approximate curve, and
subtracting the approximate curved surface from the after-chucking
main surface shape, thereby calculating an after-correction main
surface shape, and a step of selecting the substrate in which the
after-correction main surface shape has a flatness of a second
threshold value or less in the correction region.
[0039] FIG. 2 is a flowchart for explaining a mask blank substrate
manufacturing method according to the embodiment of this invention.
In this manufacturing method, first, a mask blank substrate having
a precision-polished main surface is manufactured (ST11).
[0040] In this invention, a glass substrate can be used as a mask
blank substrate. The glass substrate is not particularly limited as
long as it can be used for a mask blank. For example, there can be
cited a synthetic quartz glass, a soda-lime glass, an
aluminosilicate glass, a borosilicate glass, an alkali-free glass,
or the like.
[0041] Such a mask blank substrate can be manufactured through, for
example, a rough polishing process, a precision polishing process,
and an ultra-precision polishing process.
[0042] Then, height information in an actual measurement region of
the main surface of the mask blank substrate is obtained and, from
this height information, information of the shape of the main
surface before chucking, i.e. information of a before-chucking main
surface shape, in a cross-sectional view of the mask blank
substrate is obtained (ST12). The height information referred to
herein represents height information from a reference plane at a
plurality of measurement points set in the actual measurement
region of the main surface of the mask blank substrate. As the
actual measurement region, for example, it is possible to set a 146
mm.times.146 mm region when the size of the mask blank is 152
mm.times.152 mm. The actual measurement region is set to be a large
region including at least a virtual calculation region and a
correction region which will be described later.
[0043] The before-chucking main surface shape of the mask blank
substrate is obtained by measurement with a wavelength-shift
interferometer using a wavelength modulation laser. This
wavelength-shift interferometer calculates, as phase differences,
differences in height of a measuring surface of a mask blank
substrate from interference fringes generated by the interference
between reflected light reflected from the measuring surface and a
back surface of the mask blank substrate and a measuring apparatus
reference surface (front reference surface), detects differences in
frequency of the interference fringes, and separates the
interference fringes generated by the interference between the
reflected light reflected from the measuring surface and the back
surface of the mask blank substrate and the measuring apparatus
reference surface (front reference surface), thereby measuring the
shape of irregularities of the measuring surface.
[0044] In order to carry out a later-described simulation with high
accuracy, it is preferable that the measurement points for
obtaining the height information be set as many as possible.
However, although more accurate simulation results can be obtained
by increasing the number of the measurement points, the simulation
requires a lot of time. Therefore, it is preferable to determine
the measurement points taking these points into account. For
example, the measurement points can be set to 256.times.256
points.
[0045] Then, based on the before-chucking main surface shape of the
substrate obtained by the measurement, a flatness of the mask blank
substrate is calculated from a difference between a maximum value
and a minimum value in an actual calculation region including a
transfer region of a photomask (ST13). Further, it is judged
whether or not the flatness thus obtained is an allowable value or
less (ST14). The mask blank substrate whose flatness is greater
than the allowable value is not supplied to a subsequent process as
judged unsuccessful. Even if the flatness of an after-chucking main
surface shape of the mask blank substrate manufactured by this
manufacturing method is good, if the flatness of the
before-chucking main surface shape is poor, a problem may arise. In
the substrate in which the flatness change amount before and after
chucking is large, the substrate deformation amount is large. In a
photomask manufactured from the substrate with the large substrate
deformation amount, there is a possibility that the displacement
amount of a transfer pattern, which is formed on the substrate main
surface, before and after chucking becomes large, thus leading to a
possibility that the pattern position accuracy after chucking is
lowered. Taking this point into consideration, a selection is made
of the substrate whose flatness before chucking is the allowable
value or less. If the allowable level of the flatness before
chucking is too high, the production yield is lowered so that the
object of this invention cannot be achieved, while, if it is too
low, there arises a possibility that a photomask is manufactured
from a substrate with a large substrate deformation amount. In
consideration of the balance therebetween, the allowable value of
the flatness of the substrate before chucking is set to 0.4 .mu.m
or less. In the case of manufacturing a mask blank substrate for
use in a photomask which requires high pattern position accuracy,
since the substrate deformation amount is required to be smaller,
the allowable value of the flatness of the substrate is preferably
set to 0.3 .mu.m or less.
[0046] On the other hand, the substrate deformation amount before
and after chucking also changes depending on the size of the actual
calculation region which is a region for calculating a flatness.
There is a tendency that as the size of the actual calculation
region increases, the substrate deformation amount before and after
chucking decreases. First, it is necessary that the actual
calculation region for calculating the flatness before chucking be
a region which is equal to or smaller than the actual measurement
region as a measurement region and is larger than the virtual
calculation region. The predetermined region including the transfer
region of the photomask is determined based on an exposure
wavelength, the kind of a fine pattern (circuit pattern) to be
formed on a semiconductor wafer, and so on. When the size of the
mask blank is 152 mm square, the transfer region of the photomask
is often set to a 104 mm.times.132 mm region. Taking this into
account, the virtual calculation region can be set to a 132 mm
square region. However, if the flatness of a region outside thereof
is poor, the possibility increases that the substrate deformation
amount before and after chucking becomes large. Also taking this
into account, when the size of the mask blank (photomask) is 152 mm
square, the actual calculation region is preferably set to an at
least 142 mm square region. Further, when the measurement accuracy
of the before-chucking main surface shape and the simulation
accuracy are high, the actual calculation region is preferably set
to a 146 mm square region or a 148 mm square region.
[0047] Then, based on the obtained before-chucking main surface
shape and the shape of a mask stage, height information when the
photomask manufactured from this substrate is set in an exposure
apparatus is obtained through simulation and, from this height
information, an after-chucking main surface shape in a
cross-sectional view of the mask blank substrate is obtained
(ST15). In this simulation process, by simulating a state where the
photomask is set on the mask stage of the exposure apparatus,
height information from the reference plane is obtained at the
plurality of measurement points on the main surface of the
substrate. Herein, information necessary for obtaining, through
simulation, the height information at the plurality of measurement
points on the substrate when the photomask is set in the exposure
apparatus are the height information from the reference plane at
the plurality of measurement points on the main surface of the
substrate, which was obtained for obtaining the information of the
above-mentioned before-chucking main surface shape, and shape
information of the mask stage of the exposure apparatus including
regions where the mask stage contacts the main surface of the
substrate. According to a deflection differential equation in
mechanics of materials by the use of these information, it is
possible to obtain, through simulation, the height information from
the reference plane at the plurality of measurement points on the
main surface of the substrate when the photomask is set on the mask
stage of the exposure apparatus. Then, an after-chucking main
surface shape in a cross-sectional view of the substrate is
obtained from the obtained height information.
[0048] In this process, based on the information of the
before-chucking main surface shape of the substrate, the simulation
is carried out in which the photomask manufactured from this
substrate is chucked on the mask stage of the exposure apparatus.
Since a thin film for transfer pattern formation will be formed on
the surface of the mask blank substrate with high accuracy by
sputtering in a later process, the change in thickness of this thin
film in a main surface direction is in a range which is very small
as compared with the flatness of the substrate and thus is
ignorable. It can be said that even if the simulation is carried
out based on the before-chucking main surface shape of the mask
blank substrate, there arises no difference large enough to be
influential.
[0049] The above-mentioned deflection differential equation is
derived in the following manner, wherein a positive direction of
Z-axis is defined as the direction of gravity.
H.sub.2=H.sub.1+B.sub.1+B.sub.2-H.sub.AB
[0050] H.sub.2: height information on the substrate main surface
after chucking
[0051] H.sub.1: height information on the substrate main surface
before chucking
[0052] B.sub.1: a warp of the substrate with respect to the mask
stage as fulcrums (lever effect)
[0053] B.sub.2: a deflection of the substrate due to gravity
(approximately equal to 0.1 .mu.m: maximum value at the center of
the substrate)
[0054] H.sub.AB: an average value of height information of the
substrate in regions along the scan direction where the substrate
contacts the mask stage
[0055] The above-mentioned shape information of the mask stage may
include, in addition to the positions and regions where the mask
stage contacts the main surface of the substrate (regions each
having a slit-direction width and a scan-direction width), flatness
information of the mask stage in the above-mentioned regions
(surfaces) where the mask stage contacts the main surface of the
substrate. Further, the simulation method is not limited to the
above and a simulation using the general finite element method may
be employed.
[0056] Then, based on the after-chucking shape obtained through
simulation, a flatness of the mask blank substrate is calculated
from a difference between a maximum value and a minimum value in
the virtual calculation region including the transfer region of the
photomask (ST16). This flatness contributes to the formation of an
excellent transfer pattern upon pattern transfer using the exposure
apparatus. The virtual calculation region including the transfer
region of the photomask is determined based on an exposure
wavelength, the kind of a fine pattern (circuit pattern) to be
formed on a semiconductor wafer, and so on. When the size of the
mask blank is 152 mm square, the transfer region of the mask is
often set to a 104 mm.times.132 mm region. Taking this into
account, the virtual calculation region can be set to a 132 mm
square region. However, if the flatness of a region outside thereof
is poor, there is a possibility that the substrate deformation
amount before and after chucking becomes large. In a photomask
manufactured from such a substrate, there is a possibility that the
displacement amount of a transfer pattern formed on the substrate
main surface becomes large and thus the pattern position accuracy
after chucking is lowered. Also taking this into account, when the
size of the mask blank (photomask) is 152 mm square, the virtual
calculation region is preferably set to an at least 142 mm square
region. Further, when the measurement accuracy of the
before-chucking main surface shape and the simulation accuracy are
high, the virtual calculation region is preferably set to a 146 mm
square region or a 148 mm square region.
[0057] Then, it is judged whether or not the flatness thus obtained
is a first threshold value or less (ST17). This first threshold
value of the flatness is selected based on a flatness of an
after-correction main surface shape required for a photomask which
is manufactured from the mask blank substrate and an allowable
value of a flatness correction amount that allows the image
formation of a transfer pattern on a resist film on a semiconductor
wafer to satisfy a predetermined pattern resolution by
height-direction correction by the exposure apparatus. When the
flatness of the main surface after chucking required for the
photomask (mask blank substrate) is 0.16 .mu.m or less and the
allowable value of the flatness correction amount of the exposure
apparatus is 0.16 .mu.m, the first threshold value can be set to
0.32 .mu.m which is the sum of them. When the flatness of the main
surface after chucking required for the photomask (mask blank
substrate) is 0.08 .mu.m or less and the allowable value of the
flatness correction amount of the exposure apparatus is 0.16 .mu.m,
the first threshold value can be set to 0.24 .mu.m which is the sum
of them.
[0058] When the correction technique of an exposure apparatus is
improved so that the allowable flatness correction amount is
increased, the first threshold value can be set greater according
to the upper limit of that correction amount. For example, when the
flatness correction amount becomes greater such as 0.24 .mu.m or
0.32 .mu.m while the flatness required for the main surface after
chucking of the photomask is 0.16 .mu.m or less, the first
threshold value can be set as large as 0.40 .mu.m or 0.48 .mu.m.
When the flatness required for the main surface after chucking of
the photomask is 0.08 .mu.m or less, the first threshold value can
be set as large as 0.32 .mu.m or 0.40 .mu.m.
[0059] Conversely, when the correction technique of an exposure
apparatus is not improved while an increase in NA of the exposure
apparatus further proceeds, the focus latitude for a reduction
optical system and a semiconductor wafer decreases so that the
allowable flatness correction amount becomes smaller, and
therefore, the first threshold value should be set smaller. For
example, when the flatness correction amount becomes greater such
as 0.12 .mu.m, 0.10 .mu.m, or 0.08 .mu.m while the flatness
required for the main surface after chucking of the photomask is
0.16 .mu.m or less, the first threshold value should be set as
small as 0.28 .mu.m, 0.26 .mu.m, or 0.24 .mu.m. When the flatness
required for the main surface after chucking of the photomask is
0.08 .mu.m or less, the first threshold value should be set as
small as 0.20 .mu.m, 0.18 .mu.m, or 0.16 .mu.m.
[0060] Then, in the correction region of the after-chucking main
surface shape, shape correction is performed by calculating a first
approximate curve approximate to a cross-sectional shape along a
first direction, calculating an approximate curved surface from the
first approximate curve, and subtracting the approximate curved
surface from the after-chucking main surface shape, thereby
calculating an after-correction main surface shape (ST18). In the
case of the exposure apparatus of the type that can perform main
surface shape correction in two directions, i.e. the scan direction
and the slit direction, shape correction is performed by
calculating a first approximate curve approximate to a
cross-sectional shape along a first direction, calculating a second
approximate curve approximate to a cross-sectional shape of the
correction region along a second direction perpendicular to the
first direction, calculating an approximate curved surface from the
first approximate curve and the second approximate curve, and
subtracting the approximate curved surface from the after-chucking
main surface shape, thereby calculating an after-correction main
surface shape.
[0061] This main surface shape correction will be described using
FIGS. 3 and 4. Herein, a description will be given of the case
where the first direction is the scan direction of the exposure
apparatus, the second direction is the slit direction of the
exposure apparatus, the first approximate curve is a quartic curve,
and the second approximate curve is a quadratic curve. FIG. 3 is
diagrams for explaining main surface shape correction in the scan
direction, wherein (a) is a diagram showing positions where
cross-sectional shapes of a substrate are obtained and (b) is a
diagram showing the cross-sectional shapes of the substrate. FIG. 4
is diagrams for explaining main surface shape correction in the
slit direction, wherein (a) is a diagram showing positions where
cross-sectional shapes of a substrate are obtained and (b) is a
diagram showing the substrate shapes.
[0062] In the main surface shape correction in the scan direction
(first direction), as shown in FIG. 3(a), cross-sectional shapes of
the substrate in the scan direction are respectively obtained from
height information along right-end, middle, and left-end straight
lines Y.sub.1, each being parallel to the scan direction, in a
correction region X of an after-chucking main surface shape of a
mask blank substrate and, then, by calculating a quartic curve for
the cross-sectional shapes at the three positions by the method of
least squares, an approximate curve (first approximate curve)
Z.sub.1 in the scan direction is obtained as shown in FIG. 3(b). In
the main surface correction in the slit direction (second
direction), as shown in FIG. 4(a), cross-sectional shapes of the
substrate in the slit direction are respectively obtained from
height information along upper-end, middle, and lower-end straight
lines Y.sub.2, each being parallel to the slit direction, in a
correction region X of an after-chucking main surface shape of a
mask blank substrate and, then, by calculating a quadratic curve
for the cross-sectional shapes at the three positions by the method
of least squares, an approximate curve (second approximate curve)
Z.sub.2 in the slit direction is obtained as shown in FIG. 4(b).
Then, in the case of the exposure apparatus adapted to perform
correction only in the scan direction (first direction), an
approximate curved surface is calculated from the first approximate
curve Z.sub.1, while, in the case of the exposure apparatus adapted
to perform correction only in the slit direction (second
direction), an approximate curved surface is calculated from the
second approximate curve Z.sub.2. Then, correction is performed by
subtracting this approximate curved surface from the after-chucking
shape obtained through simulation, thereby calculating an
after-correction main surface shape.
[0063] On the other hand, in the case of the exposure apparatus
adapted to perform correction in both the scan direction (first
direction) and the slit direction (second direction), correction is
performed by calculating an approximate curved surface from the
first approximate curve Z.sub.1 and the second approximate curve
Z.sub.2 and then subtracting this approximate curved surface from
the after-chucking shape obtained through simulation, thereby
calculating an after-correction main surface shape. Alternatively,
correction is performed by subtracting an approximate curved
surface calculated from the first approximate curve Z.sub.1 and
further subtracting an approximate curved surface calculated from
the second approximate curve Z.sub.2, from the after-chucking shape
obtained through simulation, thereby calculating an
after-correction main surface shape.
[0064] This main surface shape correction simulates main surface
shape correction which is performed by the exposure apparatus
having the correction function. The after-correction main surface
shape obtained by this simulation becomes basically the same as a
substrate shape after a photomask manufactured from the substrate
subjected to this simulation is chucked in the exposure apparatus
and the main surface shape correction is performed (it does not
become completely the same due to mechanical error of the exposure
apparatus, flatness measurement error, simulation error, change in
shape due to attachment of a pellicle, etc., but this difference is
small enough not to affect judgment).
[0065] With respect to the correction region X, a transfer region,
where a transfer pattern is formed, of a photomask is determined
based on an exposure wavelength, the kind of a fine pattern
(circuit pattern) to be formed on a semiconductor wafer, and so on.
When the size of a mask blank is 152 mm square, the transfer region
of the mask is often set to a 104 mm.times.132 mm region. Taking
this into account, the correction region X is preferably set to a
132 mm square region. When higher accuracy is required for the
photomask, the correction region X is preferably set to a 142 mm
square region.
[0066] Herein, the substrate cross-sectional shapes are obtained at
the three positions in the correction region X in each of the scan
direction (first direction) and the slit direction (second
direction). On the other hand, when higher simulation accuracy is
required for the first approximate curve, the second approximate
curve, and the approximate curved surface/surfaces calculated
therefrom, each approximate curve may be calculated from substrate
cross-sectional shapes at four or more positions. Further, herein,
a quartic curve is used as the first approximate curve and a
quadratic curve is used as the second approximate curve, but not
limited thereto. A quadratic curve may be applied to the first
approximate curve while a quartic curve may be applied to the
second approximate curve. Alternatively, the first approximate
curve and the second approximate curve may both be quadratic curves
(in this case, a composite approximate curved surface becomes a
quadratic surface) or quartic curves (in this case, a composite
approximate curved surface becomes a quartic surface). It is most
preferable that a selection be made so that a correction simulation
becomes closest to the main surface shape correction function of
the exposure apparatus that actually chucks the photomask.
[0067] Then, based on the after-correction main surface shape
obtained by the main surface shape correction, a flatness of the
mask blank substrate after the correction is calculated from a
difference between a maximum value and a minimum value of height
information in the correction region including the transfer region
of the photomask, and it is judged whether or not the flatness is a
second threshold value or less (ST19).
[0068] The after-correction main surface shape is obtained by
simulating the substrate main surface shape after the photomask
(mask blank substrate) is chucked on the mask stage of the exposure
apparatus and the main surface shape correction is performed by the
correction function. That is, if this after-correction main surface
shape satisfies the flatness of the pattern transfer region
(correction region is a region including at least the pattern
transfer region) required for the photomask which is manufactured
using this mask blank substrate, such a mask blank substrate can be
judged as a successful mask blank substrate. Therefore, as the
second threshold value herein, a selection is preferably made of
the flatness of the pattern transfer region required for the
photomask which is manufactured using the mask blank. For example,
the correction region is set to a 132 mm square region and the
second threshold value is set to 0.16 .mu.m. In the case of a mask
blank substrate for use in a photomask which requires higher
accuracy, the correction region may be set to a 132 mm square
region and the second threshold value may be set to 0.08 .mu.m.
Only the mask blank substrate judged to satisfy the flatness of
this second threshold value is supplied to later-described mask
blank and photomask manufacturing processes. The above-mentioned
respective regions are preferably set with respect to the center of
the substrate main surface.
[0069] By forming at least a light-shielding film on the main
surface of the mask blank substrate judged to be a successful
product with a flatness of the second threshold value or less, a
mask blank can be obtained (ST20). As a material forming this
light-shielding film, there can be cited chromium, metal silicide,
or tantalum. Depending on the use and structure of a photomask,
another film such as an antireflection film or a phase shift film
may be appropriately formed. As a material of the antireflection
film, it is preferable to use CrO, CrON, CrOCN, or the like in the
case of a chromium-based material, MoSiON, MoSiO, MoSiN, MoSiOC, or
MoSiOCN in the case of a MoSi-based material, or TaO, TaON, TaBO,
TaBON, or the like in the case of a tantalum-based material. As a
material of the phase shift film, it is preferable to use MSiON,
MSiO, MSiN, MSiOC, MSiOCN (M: Mo, W, Ta, Zr, or the like), or the
like.
[0070] The light-shielding film can be formed by sputtering. As a
sputtering apparatus, it is possible to use a DC magnetron
sputtering apparatus, an RF magnetron sputtering apparatus, or the
like. When sputtering the light-shielding film on the mask blank
substrate, it is preferable to rotate the substrate and to dispose
a sputtering target at a position inclined by a predetermined angle
with respect to a rotation axis of the substrate, thereby forming
the light-shielding film. By such a film forming method, it is
possible to reduce in-plane variation of the light-shielding film
and thus to uniformly form the light-shielding film.
[0071] In the case of carrying out the film formation by rotating
the substrate and disposing the sputtering target at the position
inclined by the predetermined angle with respect to the rotation
axis of the substrate, the in-plane distributions of the phase
angle and the transmittance also change by the positional
relationship between the substrate and the target. The positional
relationship between the target and the substrate will be explained
with reference to FIG. 5. The offset distance (distance between the
central axis of the substrate and a straight line passing through
the center of the target and parallel to the central axis of the
substrate) is adjusted by an area in which the phase angle and
transmittance distributions are to be ensured. In general, when
such an area for ensuring the distributions is large, the required
offset distance becomes long. In this embodiment, in order to
realize a phase angle distribution within .+-.2.degree. and a
transmittance distribution within .+-.0.2% in 142 mm square of the
substrate, the offset distance is required to be about 200 mm to
350 mm and is preferably 240 mm to 280 mm. The optimal range of the
target-substrate vertical distance (T/S) changes depending on the
offset distance, but in order to realize the phase angle
distribution within .+-.2.degree. and the transmittance
distribution within .+-.0.2% in 142 mm square of the substrate, the
target-substrate vertical distance (T/S) is required to be about
200 mm to 380 mm and is preferably 210 mm to 300 mm. The target
inclination angle affects the film forming rate and, in order to
obtain a high film forming rate, the target inclination angle is
suitably 0.degree. to 45.degree. and preferably 10.degree. to
30.degree..
[0072] By patterning at least the above-mentioned light-shielding
film by photolithography and etching to form a transfer pattern, a
photomask can be manufactured (ST21). An etchant for etching is
properly changed depending on a material of a film to be
etched.
[0073] The obtained photomask is set on the mask stage of the
exposure apparatus and, using this photomask and using
photolithography with ArF excimer laser light as exposure light,
the mask pattern of the photomask is transferred to a resist film
formed on a semiconductor wafer to thereby form a required circuit
pattern on the semiconductor wafer, so that a semiconductor device
is manufactured.
[0074] Next, a description will be given of Examples which were
carried out for clarifying the effect of this invention. In the
following Examples, a description will be given of the case where a
mask blank substrate is a glass substrate.
Example 1
[0075] A predetermined number of glass substrates (about 152
mm.times.152 mm.times.6.45 mm) obtained by lapping and chamfering
synthetic quartz glass substrates were set in a double-side
polishing machine and subjected to a rough polishing process under
the following polishing conditions. After the rough polishing
process, the glass substrates were ultrasonically cleaned for
removing polishing abrasive particles adhering to the glass
substrates. The polishing conditions such as the processing
pressure, the rotational speeds of upper and lower surface plates,
and the polishing time were properly adjusted. [0076] Polishing
Liquid: cerium oxide (average particle size 2 .mu.m to 3
.mu.m)+water [0077] Polishing Pad: hard polisher (urethane pad)
[0078] Then, the predetermined number of the glass substrates after
the rough polishing were set in a double-side polishing machine and
subjected to a precision polishing process under the following
polishing conditions. After the precision polishing process, the
glass substrates were ultrasonically cleaned for removing polishing
abrasive particles adhering to the glass substrates. The polishing
conditions such as the processing pressure, the rotational speeds
of upper and lower surface plates, and the polishing time were
properly adjusted. The polishing is carried out by adjusting
various conditions so that the shape of a main surface, on the side
where a transfer pattern is to be formed, of each of the glass
substrates after the precision polishing process becomes convex at
four corners. This is because the next ultra-precision polishing
process has a feature to preferentially polish four corners of the
substrate main surface, and thus this makes it possible to suppress
edge exclusion at the four corners and to achieve a flatness of 0.4
.mu.m or less in 142 mm square of the substrate main surface.
[0079] Polishing Liquid: cerium oxide (average particle size 1
.mu.m)+water [0080] Polishing Pad: soft polisher (suede type)
[0081] Then, the predetermined number of the glass substrates after
the precision polishing were set in a double-side polishing machine
and subjected to the ultra-precision polishing process under the
following polishing conditions. After the ultra-precision polishing
process, the glass substrates were ultrasonically cleaned for
removing polishing abrasive particles adhering to the glass
substrates. The polishing conditions such as the processing
pressure, the rotational speeds of upper and lower surface plates,
and the polishing time were properly adjusted. In this
ultra-precision polishing process, there is the feature that the
four corners tend to be preferentially polished due to the
substrate shape being square. The polishing conditions are set so
that the flatness in 142 mm square of the substrate main surface
does not exceed 0.4 .mu.m while the surface roughness of the
substrate main surface becomes a predetermined roughness of 0.4 nm
or less. In this manner, the glass substrates according to this
invention were manufactured (ST11). [0082] Polishing Liquid:
colloidal silica (average particle size 100 nm)+water [0083]
Polishing Pad: super-soft polisher (suede type)
[0084] With respect to the main surface of each of the glass
substrates thus obtained, with the use of a wavelength-shift
interferometer using a wavelength modulation laser, information of
a before-chucking main surface shape (height information from a
focal plane (virtual absolute plane) calculated by the method of
least squares) was obtained at 256.times.256 measurement points in
an actual measurement region (150 mm.times.150 mm) of the main
surface (main surface where a thin film was to be formed) of the
substrate (see a before-chucking main surface shape in FIG. 6(a))
and was stored in a computer (ST12). Then, a flatness in an actual
calculation region (142 mm.times.142 mm) was obtained from the
measured height information of the before-chucking main surface
shape in the actual measurement region (ST13) and the substrate
with an allowable value (0.4 .mu.m) or less was selected (ST14). As
a result, the number of the glass substrates satisfying this
condition was 99 out of 100. From this height information, the
surface shape of the main surface of each substrate was such that
the height of the main surface was gradually reduced from its
central region toward its peripheral portion.
[0085] Then, based on the obtained information of the
before-chucking main surface shape and shape information of a mask
stage of an exposure apparatus, an after-chucking main surface
shape being information of the height from the reference plane when
the substrate was set in the exposure apparatus was calculated
through simulation at the respective measurement points using the
above-mentioned deflection differential equation (see an
after-chucking main surface shape in FIG. 6(b)) (ST15). Then, from
the result of this simulation, a difference between a maximum value
and a minimum value from the reference plane was obtained in a
virtual calculation region (142 mm.times.142 mm) including a
transfer region of a photomask, thereby calculating a flatness in
the virtual calculation region (ST16). Then, the substrate with a
flatness of a first threshold value (0.32 .mu.m) or less was
selected (ST17). As a result, the number of the substrates
satisfying this condition was 98 out of 99.
[0086] Then, main surface shape correction in the scan direction
(first direction) was performed (ST18). With respect to the glass
substrate having the after-chucking main surface shape shown in
FIG. 6(b), as shown in FIG. 7(a), cross-sectional shapes of the
substrate in the scan direction are respectively obtained from
height information along right-end, middle, and left-end straight
lines Y.sub.1, each being parallel to the scan direction, in a
correction region X of the after-chucking main surface shape of the
mask blank substrate and, then, by calculating a quartic curve for
the cross-sectional shapes at the three positions by the method of
least squares, an approximate curve (first approximate curve)
Z.sub.1 in the scan direction is obtained as shown in FIG. 7(b).
Then, an approximate curved surface shown in FIG. 7(c) is
calculated from the approximate curve Z.sub.1 and is subtracted
from the after-chucking shape obtained through simulation. An
after-correction main surface shape of the glass substrate after
subtracting the approximate curved surface Z.sub.1 is shown in FIG.
7(d).
[0087] Then, a flatness in the correction region (132 mm.times.132
mm) of the calculated after-correction main surface was calculated
and the substrate with a second threshold value (0.16 .mu.m) or
less was selected (ST19). This second threshold value was a
criterion flatness required for the mask blank substrate. As a
result, the number of the glass substrates satisfying this
condition was 96 out of 98. If the substrate satisfying the
condition of 0.16 .mu.m, equal to the second threshold value, or
less is selected based on the flatness (ST16) obtained from the
after-chucking main surface shape of the substrate calculated by
the conventional simulation (ST15), the number of the selected
substrates is 90 out of 98. Therefore, it is seen that the
production yield is largely improved by performing the main surface
shape correction (ST18).
[0088] Then, on the glass substrate thus obtained, a back-surface
antireflection layer, a light-shielding layer, and a front-surface
antireflection layer were formed in this order as a thin film
(light-shielding film) for transfer pattern formation (ST20).
Specifically, using a Cr target as a sputtering target and using a
mixed gas of Ar, CO.sub.2, N.sub.2, and He (gas flow rate ratio
Ar:CO.sub.2:N.sub.2:He=24:29:12:35) as a sputtering gas, a CrOCN
film was formed to a thickness of 39 nm as the back-surface
antireflection layer by setting the gas pressure to 0.2 Pa and the
power of the DC power supply to 1.7 kW. Then, using a Cr target as
a sputtering target and using a mixed gas of Ar, NO, and He (gas
flow rate ratio Ar:NO:He=27:18:55) as a sputtering gas, a CrON film
was formed to a thickness of 17 nm as the light-shielding layer by
setting the gas pressure to 0.1 Pa and the power of the DC power
supply to 1.7 kW. Then, using a Cr target as a sputtering target
and using a mixed gas of Ar, CO.sub.2, N.sub.2, and He (gas flow
rate ratio Ar:CO.sub.2:N.sub.2:He=21:37:11:31) as a sputtering gas,
a CrOCN film was formed to a thickness of 14 nm as the
front-surface antireflection layer by setting the gas pressure to
0.2 Pa and the power of the DC power supply to 1.8 kW. The
back-surface antireflection layer, the light-shielding layer, and
the front-surface antireflection layer formed under these
conditions had very low stress over the entire light-shielding film
and thus it was possible to suppress the change in shape of the
substrate to minimum. In this manner, a mask blank was
manufactured.
[0089] By patterning the light-shielding film of the mask blank
thus obtained into a predetermined pattern, a photomask (binary
mask) was manufactured (ST21). The obtained photomask was verified
using an exposure apparatus capable of performing main surface
shape correction at least in the scan direction. After chucking the
photomask on a mask stage of the exposure apparatus and performing
main surface shape correction in the scan direction, the pattern of
the photomask was exposed and transferred to a resist film of a
semiconductor wafer W. The CD accuracy and the pattern position
accuracy of the pattern transferred to the resist film were
verified and it was confirmed that the photomask was sufficiently
adaptable to the DRAM hp32 nm generation.
Example 2
[0090] Steps ST11 to ST17 were carried out in the same manner as in
Example 1 to select 98 glass substrates. Then, main surface shape
correction in the slit direction (second direction) was performed
(ST18). With respect to the glass substrate having the
after-chucking main surface shape shown in FIG. 6(b), as shown in
FIG. 8(a), cross-sectional shapes of the substrate in the slit
direction are respectively obtained from height information along
upper-end, middle, and lower-end straight lines Y.sub.2, each being
parallel to the slit direction, in a correction region X of the
after-chucking main surface shape of the mask blank substrate and,
then, by calculating a quadratic curve for the cross-sectional
shapes at the three positions by the method of least squares, an
approximate curve (second approximate curve) Z.sub.2 in the slit
direction is obtained as shown in FIG. 8(b). Then, an approximate
curved surface shown in FIG. 8(c) is calculated from the
approximate curve Z.sub.2 and is subtracted from the after-chucking
shape obtained through simulation. A substrate shape after
subtracting the approximate curved surface Z.sub.1 is shown in FIG.
8(d).
[0091] Then, a flatness in the correction region (132 mm.times.132
mm) of the calculated after-correction main surface was calculated
and the substrate with a second threshold value (0.16 .mu.m) or
less was selected (ST19). This second threshold value was a
criterion flatness required for the mask blank substrate. As a
result, the number of the glass substrates satisfying this
condition was 95 out of 98. If the substrate satisfying the
condition of 0.16 .mu.m, equal to the second threshold value, or
less is selected based on the flatness obtained from the
after-chucking main surface shape of the substrate (ST15)
calculated by the conventional simulation (ST14), the number of the
selected substrates is 90 out of 98. Therefore, it is seen that the
production yield is largely improved by performing the main surface
shape correction (ST18).
[0092] Then, on the glass substrate thus obtained, a phase shift
film and a light-shielding film comprising a back-surface
antireflection layer, a light-shielding layer, and a front-surface
antireflection layer were formed as a thin film for transfer
pattern formation, thereby manufacturing a mask blank (ST20).
Specifically, using a target of Mo:Si=10:90 (at % ratio) and using
a mixed gas of Ar, N.sub.2, and He (gas flow rate ratio
Ar:N.sub.2:He=5:49:46) as a sputtering gas, a MoSiN film was formed
to a thickness of 69 nm as the phase shift film by setting the gas
pressure to 0.3 Pa and the power of the DC power supply to 2.8 kW.
Then, the substrate formed with the phase shift film was
heat-treated (annealed) at 250.degree. C. for 5 minutes.
[0093] Then, the light-shielding film comprising the back-surface
antireflection layer, the light-shielding layer, and the
front-surface antireflection layer was formed. Specifically, first,
using a Cr target as a sputtering target and using a mixed gas of
Ar, CO.sub.2, N.sub.2, and He (gas flow rate ratio
Ar:CO.sub.2:N.sub.2:He=22:39:6:33) as a sputtering gas, a CrOCN
film was formed to a thickness of 30 nm as the back-surface
antireflection layer by setting the gas pressure to 0.2 Pa and the
power of the DC power supply to 1.7 kW. Then, using a Cr target as
a sputtering target and using a mixed gas of Ar and N.sub.2 (gas
flow rate ratio Ar:N.sub.2=83:17) as a sputtering gas, a CrN film
was formed to a thickness of 4 nm as the light-shielding layer by
setting the gas pressure to 0.1 Pa and the power of the DC power
supply to 1.7 kW. Then, using a Cr target as a sputtering target
and using a mixed gas of Ar, CO.sub.2, N.sub.2, and He (gas flow
rate ratio Ar:CO.sub.2:N.sub.2:He=21:37:11:31) as a sputtering gas,
a CrOCN film was formed to a thickness of 14 nm as the
front-surface antireflection layer by setting the gas pressure to
0.2 Pa and the power of the DC power supply to 1.8 kW. The
back-surface antireflection layer, the light-shielding layer, and
the front-surface antireflection layer formed under these
conditions had very low stress over the entire light-shielding film
and the phase shift film also had very low stress, and thus it was
possible to suppress the change in shape of the substrate to
minimum.
[0094] Further, by patterning the light-shielding film and the
phase shift film of the mask blank into a predetermined pattern, a
photomask (phase shift mask) was manufactured (ST21). The obtained
photomask was verified using an exposure apparatus capable of
performing main surface shape correction at least in the slit
direction. After chucking the photomask on a mask stage of the
exposure apparatus and performing main surface shape correction in
the slit direction, the pattern of the photomask was exposed and
transferred to a resist film of a semiconductor wafer W. The CD
accuracy and the pattern position accuracy of the pattern
transferred to the resist film were verified and it was confirmed
that the photomask was sufficiently adaptable to the DRAM hp32 nm
generation.
Example 3
[0095] Steps ST11 to ST17 were carried out in the same manner as in
Example 1 to select 98 glass substrates. Then, main surface shape
correction in both the scan direction (first direction) and the
slit direction (second direction) was performed (ST18). With
respect to the glass substrate having the after-chucking main
surface shape shown in FIG. 6(b), as shown in FIG. 7(a),
cross-sectional shapes of the substrate in the scan direction are
respectively obtained from height information along right-end,
middle, and left-end straight lines Y.sub.1, each being parallel to
the scan direction, in a correction region X of the after-chucking
main surface shape of the mask blank substrate and, then, by
calculating a quartic curve for the cross-sectional shapes at the
three positions by the method of least squares, an approximate
curve (first approximate curve) Z.sub.1 in the scan direction is
obtained as shown in FIG. 7(b). Further, as shown in FIG. 8(a),
cross-sectional shapes of the substrate in the slit direction are
respectively obtained from height information along upper-end,
middle, and lower-end straight lines Y.sub.2, each being parallel
to the slit direction, in the correction region X of the
after-chucking main surface shape of the mask blank substrate and,
then, by calculating a quadratic curve for the cross-sectional
shapes at the three positions by the method of least squares, an
approximate curve (second approximate curve) Z.sub.2 in the slit
direction is obtained as shown in FIG. 8(b). Then, an approximate
curved surface Z.sub.3 shown in FIG. 9(a) is calculated from the
approximate curve (first approximate curve) Z.sub.1 in the scan
direction and the approximate curve (second approximate curve)
Z.sub.2 in the slit direction and is subtracted from the
after-chucking shape obtained through simulation. A substrate shape
after subtracting the approximate curved surface Z.sub.3 is shown
in FIG. 9(b).
[0096] Then, a flatness in the correction region (132 mm.times.132
mm) of the calculated after-correction main surface was calculated
and the substrate with a second threshold value (0.16 .mu.m) or
less was selected (ST19). This second threshold value was a
criterion flatness required for the mask blank substrate. As a
result, the number of the glass substrates satisfying this
condition was 97 out of 98. If the substrate satisfying the
condition of 0.16 .mu.m, equal to the second threshold value, or
less is selected based on the flatness obtained from the
after-chucking main surface shape of the substrate (ST15)
calculated by the conventional simulation (ST14), the number of the
selected substrates is 90 out of 98. Therefore, it is seen that the
production yield is largely improved by performing the main surface
shape correction (ST18).
[0097] Then, on the glass substrate thus obtained, a MoSiON film
(back-surface antireflection layer), a MoSi film (light-shielding
layer), and a MoSiON film (antireflection layer) were formed as a
thin film (light-shielding film) for transfer pattern formation,
thereby manufacturing a mask blank (ST20).
[0098] Specifically, using a target of Mo:Si=21:79 (at % ratio) and
using Ar, O.sub.2, N.sub.2, and He (gas flow rate ratio
Ar:O.sub.2:N.sub.2:He=5:4:49:42) at a sputtering gas pressure of
0.2 Pa, a film made of molybdenum, silicon, oxygen, and nitrogen
(MoSiON film) was formed to a thickness of 7 nm by setting the
power of the DC power supply to 3.0 kW. Then, using the same target
and using Ar and He (gas flow rate ratio Ar:He=20:120) at a
sputtering gas pressure of 0.3 Pa, a film made of molybdenum and
silicon (MoSi film: at % ratio of Mo and Si in the film was about
21:79) was formed to a thickness of 30 nm by setting the power of
the DC power supply to 2.0 kW. Then, using a target of Mo:Si=4:96
(at % ratio) and using Ar, O.sub.2, N.sub.2, and He (gas flow rate
ratio Ar:O.sub.2:N.sub.2:He=6:5:11:16) at a sputtering gas pressure
of 0.1 Pa, a film made of molybdenum, silicon, oxygen, and nitrogen
(MoSiON film) was formed to a thickness of 15 nm by setting the
power of the DC power supply to 3.0 kW. The total thickness of the
thin film (light-shielding film) was set to 52 nm. The back-surface
antireflection layer, the light-shielding layer, and the
front-surface antireflection layer formed under these conditions
had very low stress over the entire light-shielding film and thus
it was possible to suppress the change in shape of the substrate to
minimum.
[0099] Further, by patterning the light-shielding film and the
antireflection film of the mask blank into a predetermined pattern,
a photomask (binary mask) was manufactured (ST21). The obtained
photomask was verified using an exposure apparatus capable of
performing main surface shape correction in the scan direction and
in the slit direction. After chucking the photomask on a mask stage
of the exposure apparatus and performing main surface shape
correction in the scan direction and in the slit direction, the
pattern of the photomask was exposed and transferred to a resist
film of a semiconductor wafer W. The CD accuracy and the pattern
position accuracy of the pattern transferred to the resist film
were verified and it was confirmed that the photomask was
sufficiently adaptable to the DRAM hp32 nm generation.
Example 4
[0100] Steps ST11 to ST17 were carried out in the same manner as in
Example 1 to select 98 glass substrates and then main surface shape
correction in the scan direction (first direction) was performed
(ST18). Then, a flatness in a correction region (132 mm.times.132
mm) of a calculated after-correction main surface was calculated
and the substrate with a second threshold value (0.08 .mu.m) or
less was selected (ST19). This second threshold value was a
criterion flatness required for the mask blank substrate. As a
result, the number of the glass substrates satisfying this
condition was 92 out of 98. If the substrate satisfying the
condition of 0.08 .mu.m, equal to the second threshold value, or
less is selected based on the flatness obtained from the
after-chucking main surface shape of the substrate (ST15)
calculated by the conventional simulation (ST14), the number of the
selected substrates is 84 out of 98. Therefore, it is seen that the
production yield is largely improved by performing the main surface
shape correction (ST18).
[0101] Then, on the glass substrate thus obtained, a
light-shielding film comprising a back-surface antireflection
layer, a light-shielding layer, and a front-surface antireflection
layer was formed as a thin film for transfer pattern formation in
the same manner as in Example 1, thereby manufacturing a mask blank
(ST20). Further, by patterning the light-shielding film of the mask
blank into a predetermined pattern, a photomask (binary mask) was
manufactured (ST21). The obtained photomask was verified using an
exposure apparatus capable of performing main surface shape
correction at least in the scan direction. After chucking the
photomask on a mask stage of the exposure apparatus and performing
main surface shape correction in the scan direction, the pattern of
the photomask was exposed and transferred to a resist film of a
semiconductor wafer W. The CD accuracy and the pattern position
accuracy of the pattern transferred to the resist film were
verified and it was confirmed that the photomask was sufficiently
adaptable to the DRAM hp22 nm generation.
Example 5
[0102] Steps ST11 to ST17 were carried out in the same manner as in
Example 2 to select 98 glass substrates and then main surface shape
correction in the slit direction (second direction) was performed
(ST18). Then, a flatness in a correction region (132 mm.times.132
mm) of a calculated after-correction main surface was calculated
and the substrate with a second threshold value (0.08 .mu.m) or
less was selected (ST19). This second threshold value was a
criterion flatness required for the mask blank substrate. As a
result, the number of the glass substrates satisfying this
condition was 91 out of 98. If the substrate satisfying the
condition of 0.08 .mu.m, equal to the second threshold value, or
less is selected based on the flatness obtained from the
after-chucking main surface shape of the substrate (ST15)
calculated by the conventional simulation (ST14), the number of the
selected substrates is 84 out of 98. Therefore, it is seen that the
production yield is largely improved by performing the main surface
shape correction (ST18).
[0103] Then, on the glass substrate thus obtained, a phase shift
film and a light-shielding film comprising a back-surface
antireflection layer, a light-shielding layer, and a front-surface
antireflection layer were formed as a thin film for transfer
pattern formation in the same manner as in Example 2, thereby
manufacturing a mask blank (ST20). Further, by patterning the
light-shielding film and the phase shift film of the mask blank
into a predetermined pattern, a photomask (phase shift mask) was
manufactured (ST21). The obtained photomask was verified using an
exposure apparatus capable of performing main surface shape
correction at least in the slit direction. After chucking the
photomask on a mask stage of the exposure apparatus and performing
main surface shape correction in the slit direction, the pattern of
the photomask was exposed and transferred to a resist film of a
semiconductor wafer W. The CD accuracy and the pattern position
accuracy of the pattern transferred to the resist film were
verified and it was confirmed that the photomask was sufficiently
adaptable to the DRAM hp22 nm generation.
Example 6
[0104] Steps ST11 to ST17 were carried out in the same manner as in
Example 3 to select 98 glass substrates and then main surface shape
correction in the scan direction (first direction) and in the slit
direction (second direction) was performed (ST18). Then, a flatness
in a correction region (132 mm.times.132 mm) of a calculated
after-correction main surface was calculated and the substrate with
a second threshold value (0.08 .mu.m) or less was selected (ST19).
This second threshold value was a criterion flatness required for
the mask blank substrate. As a result, the number of the glass
substrates satisfying this condition was 93 out of 98. If the
substrate satisfying the condition of 0.08 .mu.m, equal to the
second threshold value, or less is selected based on the flatness
obtained from the after-chucking main surface shape of the
substrate (ST15) calculated by the conventional simulation (ST14),
the number of the selected substrates is 84 out of 98. Therefore,
it is seen that the production yield is largely improved by
performing the main surface shape correction (ST18).
[0105] Then, on the glass substrate thus obtained, a
light-shielding film comprising a back-surface antireflection
layer, a light-shielding layer, and a front-surface antireflection
layer was formed as a thin film for transfer pattern formation in
the same manner as in Example 3, thereby manufacturing a mask blank
(ST20). Further, by patterning the light-shielding film of the mask
blank into a predetermined pattern, a photomask (binary mask) was
manufactured (ST21). The obtained photomask was verified using an
exposure apparatus capable of performing main surface shape
correction in both the scan direction and the slit direction. After
chucking the photomask on a mask stage of the exposure apparatus
and performing main surface shape correction in the scan direction
and in the slit direction, the pattern of the photomask was exposed
and transferred to a resist film of a semiconductor wafer W. The CD
accuracy and the pattern position accuracy of the pattern
transferred to the resist film were verified and it was confirmed
that the photomask was sufficiently adaptable to the DRAM hp22 nm
generation.
[0106] This invention is not limited to the above-mentioned
embodiment and can be carried out by appropriately changing it. For
example, the materials, the sizes, the processing sequences, and so
on in the above-mentioned embodiment are only examples and this
invention can be carried out by changing them in various ways
within a range capable of exhibiting the effect of this invention.
Other than that, this invention can be carried out with appropriate
changes within a range not departing from the object of this
invention.
[0107] This application claims priority from Japanese Patent
Application No. 2009-074997, filed on Mar. 25, 2009, the disclosure
of which is incorporated herein in its entirety by reference.
DESCRIPTION OF SYMBOLS
[0108] 1 mask stage [0109] 1a chuck [0110] 2 photomask [0111] 3
slit member [0112] 3a slit [0113] 4 light source [0114] 5
illumination optical system [0115] 6 reduction optical system
[0116] 7 wafer stage [0117] W semiconductor wafer
* * * * *